Labels match but Cluster Autoscaler says "are not similar, labels do not match" when trying to balance similar node groups.

nicksecurity commented 4 months ago

Which component are you using?: cluster-autoscaler

What version of the component are you using?: Component version: 1.28.5

What k8s version are you using (kubectl version)?:

kubectl version Output

$ kubectl version
Client Version: v1.28.3
Kustomize Version: v5.0.4-0.20230601165947-6ce0bf390ce3
Server Version: v1.28.9-eks-036c24b

What environment is this in?: AWS, EKS using EC2

What did you expect to happen?: When new nodes are added I expect them to be balanced on to 3 similar node groups, which have the same labels.

What happened instead?: I have 3 node groups, one per AZ, the new nodes are only added to 1 node group.

The error says the labels are different, but I've checked them all, and except for a couple I've excluded, they are similar.

I0621 14:19:48.338960 1 compare_nodegroups.go:157] nodes template-node-for-eks-nodegroup-4-128-1ec80b90 and template-node-for-eks-nodegroup-3-128-f0c80b8d are not similar, labels do not match

How to reproduce it (as minimally and precisely as possible): I scale up the pods so it will add several new nodes and then check to see which nodegroup it added them to.

Anything else we need to know?: No

adrianmoisey commented 4 months ago

/area cluster-autoscaler

k8s-triage-robot commented 1 month ago

The Kubernetes project currently lacks enough contributors to adequately respond to all issues.

This bot triages un-triaged issues according to the following rules:

After 90d of inactivity, lifecycle/stale is applied
After 30d of inactivity since lifecycle/stale was applied, lifecycle/rotten is applied
After 30d of inactivity since lifecycle/rotten was applied, the issue is closed

You can:

Mark this issue as fresh with /remove-lifecycle stale
Close this issue with /close
Offer to help out with Issue Triage

Please send feedback to sig-contributor-experience at kubernetes/community.

/lifecycle stale

k8s-triage-robot commented 3 weeks ago

The Kubernetes project currently lacks enough active contributors to adequately respond to all issues.

This bot triages un-triaged issues according to the following rules:

After 90d of inactivity, lifecycle/stale is applied
After 30d of inactivity since lifecycle/stale was applied, lifecycle/rotten is applied
After 30d of inactivity since lifecycle/rotten was applied, the issue is closed

You can:

Mark this issue as fresh with /remove-lifecycle rotten
Close this issue with /close
Offer to help out with Issue Triage

Please send feedback to sig-contributor-experience at kubernetes/community.

/lifecycle rotten

jbilliau-rcd commented 2 weeks ago

Having same issue when testing out cluster-autoscaler 1.29.4....I have 3 node groups, one per AZ, and spun up 20 pods....it increased only ONE node group by 11?

ster-autoscaler-86c94b64cd-mv5mm aws-cluster-autoscaler I1031 14:33:48.182407       1 klogx.go:87] Pod beta-whale/beta-whale-5fb675658c-n9kls is unschedulable
cluster-autoscaler-86c94b64cd-mv5mm aws-cluster-autoscaler I1031 14:33:48.182410       1 klogx.go:87] Pod beta-whale/beta-whale-5fb675658c-hmfr7 is unschedulable
cluster-autoscaler-86c94b64cd-mv5mm aws-cluster-autoscaler I1031 14:33:48.182412       1 klogx.go:87] Pod beta-whale/beta-whale-5fb675658c-zx4gc is unschedulable
cluster-autoscaler-86c94b64cd-mv5mm aws-cluster-autoscaler I1031 14:33:48.182414       1 klogx.go:87] Pod beta-whale/beta-whale-5fb675658c-nqr45 is unschedulable
cluster-autoscaler-86c94b64cd-mv5mm aws-cluster-autoscaler I1031 14:33:48.182416       1 klogx.go:87] Pod beta-whale/beta-whale-5fb675658c-krgdk is unschedulable
cluster-autoscaler-86c94b64cd-mv5mm aws-cluster-autoscaler I1031 14:33:48.182423       1 klogx.go:87] Pod kube-system/overprovisioning-757f8f8fbc-9bftr is unschedulable
cluster-autoscaler-86c94b64cd-mv5mm aws-cluster-autoscaler I1031 14:33:48.182425       1 klogx.go:87] 1 other pods are also unschedulable
cluster-autoscaler-86c94b64cd-mv5mm aws-cluster-autoscaler I1031 14:33:48.184241       1 orchestrator.go:108] Upcoming 0 nodes
cluster-autoscaler-86c94b64cd-mv5mm aws-cluster-autoscaler I1031 14:33:48.184397       1 compare_nodegroups.go:157] nodes template-node-for-eks-spot-jason20241030133218199100000003-02c96ed0-43e2-76e6-7f01-f1c22b7851f6-1335741201004455891 and template-node-for-eks-spot-jason20241030133218207200000005-dac96ed0-43e6-8b44-fa5d-8ffe23060302-574218763908410739 are not similar, labels do not match
cluster-autoscaler-86c94b64cd-mv5mm aws-cluster-autoscaler I1031 14:33:48.184413       1 compare_nodegroups.go:157] nodes template-node-for-eks-spot-jason20241030133218199100000003-02c96ed0-43e2-76e6-7f01-f1c22b7851f6-1335741201004455891 and template-node-for-eks-spot-jason20241030133218211400000007-cec96ed0-43eb-c9e9-8f1c-1d9adaa3c01b-1772456504083316692 are not similar, labels do not match
cluster-autoscaler-86c94b64cd-mv5mm aws-cluster-autoscaler I1031 14:33:48.185704       1 compare_nodegroups.go:157] nodes template-node-for-eks-spot-jason20241030133218207200000005-dac96ed0-43e6-8b44-fa5d-8ffe23060302-574218763908410739 and template-node-for-eks-spot-jason20241030133218211400000007-cec96ed0-43eb-c9e9-8f1c-1d9adaa3c01b-1772456504083316692 are not similar, labels do not match
cluster-autoscaler-86c94b64cd-mv5mm aws-cluster-autoscaler I1031 14:33:48.185720       1 compare_nodegroups.go:157] nodes template-node-for-eks-spot-jason20241030133218207200000005-dac96ed0-43e6-8b44-fa5d-8ffe23060302-574218763908410739 and template-node-for-eks-spot-jason20241030133218199100000003-02c96ed0-43e2-76e6-7f01-f1c22b7851f6-1335741201004455891 are not similar, labels do not match
cluster-autoscaler-86c94b64cd-mv5mm aws-cluster-autoscaler I1031 14:33:48.186994       1 compare_nodegroups.go:157] nodes template-node-for-eks-spot-jason20241030133218211400000007-cec96ed0-43eb-c9e9-8f1c-1d9adaa3c01b-1772456504083316692 and template-node-for-eks-spot-jason20241030133218199100000003-02c96ed0-43e2-76e6-7f01-f1c22b7851f6-1335741201004455891 are not similar, labels do not match
cluster-autoscaler-86c94b64cd-mv5mm aws-cluster-autoscaler I1031 14:33:48.187009       1 compare_nodegroups.go:157] nodes template-node-for-eks-spot-jason20241030133218211400000007-cec96ed0-43eb-c9e9-8f1c-1d9adaa3c01b-1772456504083316692 and template-node-for-eks-spot-jason20241030133218207200000005-dac96ed0-43e6-8b44-fa5d-8ffe23060302-574218763908410739 are not similar, labels do not match
cluster-autoscaler-86c94b64cd-mv5mm aws-cluster-autoscaler I1031 14:33:48.188273       1 priority.go:114] Successfully loaded priority configuration from configmap.
cluster-autoscaler-86c94b64cd-mv5mm aws-cluster-autoscaler I1031 14:33:48.188285       1 priority.go:163] priority expander: eks-spot-jason20241030133218199100000003-02c96ed0-43e2-76e6-7f01-f1c22b7851f6 chosen as the highest available
cluster-autoscaler-86c94b64cd-mv5mm aws-cluster-autoscaler I1031 14:33:48.188288       1 priority.go:163] priority expander: eks-spot-jason20241030133218207200000005-dac96ed0-43e6-8b44-fa5d-8ffe23060302 chosen as the highest available
cluster-autoscaler-86c94b64cd-mv5mm aws-cluster-autoscaler I1031 14:33:48.188291       1 priority.go:163] priority expander: eks-spot-jason20241030133218211400000007-cec96ed0-43eb-c9e9-8f1c-1d9adaa3c01b chosen as the highest available
cluster-autoscaler-86c94b64cd-mv5mm aws-cluster-autoscaler I1031 14:33:48.188298       1 orchestrator.go:181] Best option to resize: eks-spot-jason20241030133218199100000003-02c96ed0-43e2-76e6-7f01-f1c22b7851f6
cluster-autoscaler-86c94b64cd-mv5mm aws-cluster-autoscaler I1031 14:33:48.188307       1 orchestrator.go:185] Estimated 11 nodes needed in eks-spot-jason20241030133218199100000003-02c96ed0-43e2-76e6-7f01-f1c22b7851f6
cluster-autoscaler-86c94b64cd-mv5mm aws-cluster-autoscaler I1031 14:33:48.188323       1 compare_nodegroups.go:157] nodes template-node-for-eks-spot-jason20241030133218199100000003-02c96ed0-43e2-76e6-7f01-f1c22b7851f6-1335741201004455891 and template-node-for-eks-spot-jason20241030133218207200000005-dac96ed0-43e6-8b44-fa5d-8ffe23060302-574218763908410739 are not similar, labels do not match
cluster-autoscaler-86c94b64cd-mv5mm aws-cluster-autoscaler I1031 14:33:48.188335       1 compare_nodegroups.go:157] nodes template-node-for-eks-spot-jason20241030133218199100000003-02c96ed0-43e2-76e6-7f01-f1c22b7851f6-1335741201004455891 and template-node-for-eks-spot-jason20241030133218211400000007-cec96ed0-43eb-c9e9-8f1c-1d9adaa3c01b-1772456504083316692 are not similar, labels do not match
cluster-autoscaler-86c94b64cd-mv5mm aws-cluster-autoscaler I1031 14:33:48.188342       1 orchestrator.go:249] No similar node groups found
cluster-autoscaler-86c94b64cd-mv5mm aws-cluster-autoscaler I1031 14:33:48.188358       1 orchestrator.go:291] Final scale-up plan: [{eks-spot-jason20241030133218199100000003-02c96ed0-43e2-76e6-7f01-f1c22b7851f6 2->13 (max: 50)}]
cluster-autoscaler-86c94b64cd-mv5mm aws-cluster-autoscaler I1031 14:33:48.188385       1 executor.go:147] Scale-up: setting group eks-spot-jason20241030133218199100000003-02c96ed0-43e2-76e6-7f01-f1c22b7851f6 size to 13
cluster-autoscaler-86c94b64cd-mv5mm aws-cluster-autoscaler I1031 14:33:48.188409       1 auto_scaling_groups.go:267] Setting asg eks-spot-jason20241030133218199100000003-02c96ed0-43e2-76e6-7f01-f1c22b7851f6 size to 13

jbilliau-rcd commented 2 weeks ago

Hmmm never mind, just fixed by switching to --balance-labels.....balance-ignore-labels seems overly complicated and is prone to breaking if labels change on nodes. @nicksecurity see this PR here - https://github.com/kubernetes/autoscaler/pull/4174

cluster-autoscaler-759cb97bf8-nrdjc aws-cluster-autoscaler I1031 14:59:06.230480       1 priority.go:163] priority expander: eks-spot-jason20241030133218199100000003-02c96ed0-43e2-76e6-7f01-f1c22b7851f6 chosen as the highest available
cluster-autoscaler-759cb97bf8-nrdjc aws-cluster-autoscaler I1031 14:59:06.230484       1 priority.go:163] priority expander: eks-spot-jason20241030133218207200000005-dac96ed0-43e6-8b44-fa5d-8ffe23060302 chosen as the highest available
cluster-autoscaler-759cb97bf8-nrdjc aws-cluster-autoscaler I1031 14:59:06.230486       1 priority.go:163] priority expander: eks-spot-jason20241030133218211400000007-cec96ed0-43eb-c9e9-8f1c-1d9adaa3c01b chosen as the highest available
cluster-autoscaler-759cb97bf8-nrdjc aws-cluster-autoscaler I1031 14:59:06.230493       1 orchestrator.go:181] Best option to resize: eks-spot-jason20241030133218207200000005-dac96ed0-43e6-8b44-fa5d-8ffe23060302
cluster-autoscaler-759cb97bf8-nrdjc aws-cluster-autoscaler I1031 14:59:06.230498       1 orchestrator.go:185] Estimated 8 nodes needed in eks-spot-jason20241030133218207200000005-dac96ed0-43e6-8b44-fa5d-8ffe23060302
cluster-autoscaler-759cb97bf8-nrdjc aws-cluster-autoscaler I1031 14:59:06.230520       1 orchestrator.go:246] Found 2 similar node groups: [eks-spot-jason20241030133218211400000007-cec96ed0-43eb-c9e9-8f1c-1d9adaa3c01b eks-spot-jason20241030133218199100000003-02c96ed0-43e2-76e6-7f01-f1c22b7851f6]
cluster-autoscaler-759cb97bf8-nrdjc aws-cluster-autoscaler I1031 14:59:06.230548       1 orchestrator.go:281] Splitting scale-up between 3 similar node groups: {eks-spot-jason20241030133218207200000005-dac96ed0-43e6-8b44-fa5d-8ffe23060302, eks-spot-jason20241030133218211400000007-cec96ed0-43eb-c9e9-8f1c-1d9adaa3c01b, eks-spot-jason20241030133218199100000003-02c96ed0-43e2-76e6-7f01-f1c22b7851f6}
cluster-autoscaler-759cb97bf8-nrdjc aws-cluster-autoscaler I1031 14:59:06.230567       1 orchestrator.go:291] Final scale-up plan: [{eks-spot-jason20241030133218211400000007-cec96ed0-43eb-c9e9-8f1c-1d9adaa3c01b 1->5 (max: 50)} {eks-spot-jason20241030133218207200000005-dac96ed0-43e6-8b44-fa5d-8ffe23060302 2->5 (max: 50)} {eks-spot-jason20241030133218199100000003-02c96ed0-43e2-76e6-7f01-f1c22b7851f6 3->4 (max: 50)}]
cluster-autoscaler-759cb97bf8-nrdjc aws-cluster-autoscaler I1031 14:59:06.230583       1 executor.go:147] Scale-up: setting group eks-spot-jason20241030133218211400000007-cec96ed0-43eb-c9e9-8f1c-1d9adaa3c01b size to 5
cluster-autoscaler-759cb97bf8-nrdjc aws-cluster-autoscaler I1031 14:59:06.230606       1 auto_scaling_groups.go:267] Setting asg eks-spot-jason20241030133218211400000007-cec96ed0-43eb-c9e9-8f1c-1d9adaa3c01b size to 5
cluster-autoscaler-759cb97bf8-nrdjc aws-cluster-autoscaler I1031 14:59:06.230830       1 event_sink_logging_wrapper.go:48] Event(v1.ObjectReference{Kind:"ConfigMap", Namespace:"kube-system", Name:"cluster-autoscaler-status", UID:"1cd02fde-5df3-4755-a3be-880101f4f685", APIVersion:"v1", ResourceVersion:"110291641", FieldPath:""}): type: 'Normal' reason: 'ScaledUpGroup' Scale-up: setting group eks-spot-jason20241030133218211400000007-cec96ed0-43eb-c9e9-8f1c-1d9adaa3c01b size to 5 instead of 1 (max: 50)
cluster-autoscaler-759cb97bf8-nrdjc aws-cluster-autoscaler I1031 14:59:06.353381       1 executor.go:147] Scale-up: setting group eks-spot-jason20241030133218207200000005-dac96ed0-43e6-8b44-fa5d-8ffe23060302 size to 5
cluster-autoscaler-759cb97bf8-nrdjc aws-cluster-autoscaler I1031 14:59:06.353413       1 auto_scaling_groups.go:267] Setting asg eks-spot-jason20241030133218207200000005-dac96ed0-43e6-8b44-fa5d-8ffe23060302 size to 5
cluster-autoscaler-759cb97bf8-nrdjc aws-cluster-autoscaler I1031 14:59:06.353488       1 event_sink_logging_wrapper.go:48] Event(v1.ObjectReference{Kind:"ConfigMap", Namespace:"kube-system", Name:"cluster-autoscaler-status", UID:"1cd02fde-5df3-4755-a3be-880101f4f685", APIVersion:"v1", ResourceVersion:"110291641", FieldPath:""}): type: 'Normal' reason: 'ScaledUpGroup' Scale-up: group eks-spot-jason20241030133218211400000007-cec96ed0-43eb-c9e9-8f1c-1d9adaa3c01b size set to 5 instead of 1 (max: 50)
cluster-autoscaler-759cb97bf8-nrdjc aws-cluster-autoscaler I1031 14:59:06.367471       1 event_sink_logging_wrapper.go:48] Event(v1.ObjectReference{Kind:"ConfigMap", Namespace:"kube-system", Name:"cluster-autoscaler-status", UID:"1cd02fde-5df3-4755-a3be-880101f4f685", APIVersion:"v1", ResourceVersion:"110291641", FieldPath:""}): type: 'Normal' reason: 'ScaledUpGroup' Scale-up: setting group eks-spot-jason20241030133218207200000005-dac96ed0-43e6-8b44-fa5d-8ffe23060302 size to 5 instead of 2 (max: 50)
cluster-autoscaler-759cb97bf8-nrdjc aws-cluster-autoscaler I1031 14:59:06.434135       1 executor.go:147] Scale-up: setting group eks-spot-jason20241030133218199100000003-02c96ed0-43e2-76e6-7f01-f1c22b7851f6 size to 4
cluster-autoscaler-759cb97bf8-nrdjc aws-cluster-autoscaler I1031 14:59:06.434161       1 auto_scaling_groups.go:267] Setting asg eks-spot-jason20241030133218199100000003-02c96ed0-43e2-76e6-7f01-f1c22b7851f6 size to 4
cluster-autoscaler-759cb97bf8-nrdjc aws-cluster-autoscaler I1031 14:59:06.434240       1 event_sink_logging_wrapper.go:48] Event(v1.ObjectReference{Kind:"ConfigMap", Namespace:"kube-system", Name:"cluster-autoscaler-status", UID:"1cd02fde-5df3-4755-a3be-880101f4f685", APIVersion:"v1", ResourceVersion:"110291641", FieldPath:""}): type: 'Normal' reason: 'ScaledUpGroup' Scale-up: group eks-spot-jason20241030133218207200000005-dac96ed0-43e6-8b44-fa5d-8ffe23060302 size set to 5 instead of 2 (max: 50)

kubernetes / autoscaler

Labels match but Cluster Autoscaler says "are not similar, labels do not match" when trying to balance similar node groups. #6954